association between salt and hypertension in rural and urban...

402 Asia Pac J Clin Nutr 2016;25(2):402-413

Original Article Association between salt and hypertension in rural and urban populations of low to middle income countries: a systematic review and meta-analysis of population based studies Asvini K Subasinghe PhD1, Simin Arabshahi PhD1, Doreen Busingye PhD1, Roger G Evans PhD2, Karen Z Walker PhD3, Michaela A Riddell PhD1, Amanda G Thrift PhD1,4 1Department of Medicine, School of Clinical Sciences at Monash Health, Monash University, Melbourne, Australia 2Department of Physiology, Monash University, Melbourne, Australia 3Department of Nutrition and Dietetics, Monash University, Melbourne, Australia 4Florey Neuroscience Institutes, Heidelberg, Victoria, Australia

Background and Objectives: The prevalence of hypertension, the greatest contributor to mortality globally, is increasing in low-and-middle income countries (LMICs). In urban regions of LMICs, excessive salt intake is as-sociated with increased risk of hypertension. We aimed to determine whether this is the case in rural regions as well. Methods and Study Design: We performed a meta-analysis of studies in rural and urban areas of LMICs in which the association of salt and hypertension were assessed using multivariable models. Results: We identified 18 studies with a total of 134,916 participants. The prevalence of high salt intake ranged from 21.3% to 89.5% in rural and urban populations. When salt was analysed as a continuous variable, a greater impact of salt on hyper-tension was found in urban (n=4) (pooled effect size (ES) 1.42, 95% CI 1.19, 1.69) than in rural populations (n=4) (pooled ES 1.07, 95% CI 1.04, 1.10, p for difference <0.001). In studies where salt was analysed continuously, a greater impact of salt on hypertension was observed in lean rural populations (BMI <23 kg/m2) than in non-lean rural populations (BMI ≥23 kg/m2, p for difference <0.001). Conclusions: The prevalence of high salt intake is similar in rural and urban regions. Excessive salt intake has a greater impact on the prevalence of hypertension in urban than rural regions. BMI appears to modify the relationship between salt and hypertension in rural popula-tions.

Key Words: hypertension, salt, rural, urban, meta-analysis INTRODUCTION Approximately two thirds of the 972 million people with hypertension live in low-and-middle income countries (LMICs).1-3 The prevalence of hypertension in LMICs is increasing rapidly,1-3 due in large part to economic and demographic transitions.4-6 These transitions have led to the adoption of westernised lifestyles and the increased prevalence of smoking, alcohol, sedentary activity, obesi-ty, poor dietary habits, and other risk factors for chronic disease.7

There is compelling evidence that high salt intake is an important driver of hypertension.8-14 Investigators in the landmark INTERSALT study, which included more than 10,000 adults in 32 countries, found a quasi-linear rela-tionship between salt-intake and blood pressure.15 High salt intake was only observed in one of the four IN-TERSALT cohorts from non-urban regions, and this co-hort may not have been representative of non-urban popu-lations subjected to high salt intake because the popula-tion centre was located nearly 4000 metres above sea

level.16 This highlights the limited information regarding the relationship between salt intake and hypertension in rural populations of LMICs compared with urban regions.

A noteworthy observation in the INTERSALT study was that the relationship between sodium and blood pres-sure attenuated when BMI was introduced into multiple regression models.17 An explanation for this attenuation is that the blood pressure of obese individuals is sensitive to the effects of sodium, and results in an increased activity of the sympathetic nervous system.18 Even though there is limited evidence for this relationship in populations of

Corresponding Author: Dr Asvini K Subasinghe, Department of Medicine, School of Clinical Sciences at Monash Health, Monash University Level 1, 43-51 Kanooka Grove Clayton VIC 3168, Melbourne, Australia. Tel: +613 95947572; Fax: +613 9902 4245 Email: [email protected] Manuscript received 07 February 2015. Initial review completed and accepted 26 April 2015. doi: 10.6133/apjcn.2016.25.2.25

Salt and hypertension in rural and urban populations 403

LMICs, obesity related salt-sensitivity may underscore the sodium-blood pressure relationship in LMICs as well. If so, obesity related salt-sensitivity might be expected to be more evident in urban regions, as the prevalence of obesity is greater than in rural regions.19 There is evi-dence that salt intake is high in LMICs.20,21,14 In urban populations, excessive salt intake is attributed to a nutri-tional shift from cereal and millet based diets to a diet comprising processed foods with added sugars, fats, and sodium.22 In rural populations, salt is mostly derived from its addition during the preparation of meals, as it is the cheapest form of flavour.23,24 The lack of processed foods in rural populations might lead one to expect that salt intake would be less in rural areas than in urban areas.

We undertook a systematic review and meta-analysis to determine whether (1) the prevalence of excessive salt intake, and (2) the relationship between salt intake and the risk of hypertension, differs between rural and urban populations of LMICs. We also aimed to determine whether BMI modified the association between salt and hypertension in these populations. We hypothesised that (1) salt intake in rural areas of LMICs is less than that in urban areas, (2) the contribution of salt intake to the risk of hypertension is greater in urban than in rural regions (3) The impact of BMI on salt intake and hypertension differs between these two settings. METHODS Search strategy and selection criteria We searched three databases (PubMED, Web of Science (WOS), and Scopus) until July 2014. Search terms used were “poverty areas” or “vulnerable population*” or “ru-ral population*” or “rural health” or “developing coun-tries” or “rural” or “disadvantaged areas” or “disadvan-taged population*” or “urban” or “urban population” and “low-income countries” or “middle-income countries” or “low income countries” or “low AND middle-income countries” and (“risk factor(s)” or “determinant(s)” or “predictor(s)” or “blood pressure determination(s)” or “nutritional status” or “nutrition” or “salt” or “sodium” or “sodium/urine*” or “DASH diet”) and (“hypertension” or “high blood pressure” or “blood pressure”) or “WHO steps”). Each search included publications written in Eng-lish and other languages.

We included population-based studies (with either a case-control, cross-sectional or longitudinal design) in which the association between salt and hypertension was assessed in adults using multivariable analysis. We ex-cluded studies in which only univariable associations be-tween potential risk factors and hypertension were as-sessed, in which only genetic analysis was undertaken, or which were limited to investigation of only younger (<24 years) or older (≥65 years) participants.

Classifications of countries developed by the World Bank25 were used to define the LMICs. Rural and urban populations were classified according to the reports by the authors of each published article. Where there were multiple reports from one study, the most recent public-ation was included.

The following information was extracted from each study included in the analysis: study type; region and country; year/s of data collection (cross-sectional studies);

sample size; age-range and mean age of participants; def-inition of hypertension used; mean and standard devia-tion (SD) of arterial blood pressure (BP); prevalence of hypertension; categories of salt intake and sodium intake; point estimates (odds ratios (ORs)) and 95% confidence intervals (CIs), and variables for which statistical adjust-ments were performed. When estimates from multiple models were reported, we used the model that provided adjustment for the greatest number of variables. Definitions Only studies in which both systolic blood pressure (SBP) and diastolic blood pressure (DBP) were used to define hypertension were considered for inclusion in the meta-analysis. Hypertension was defined as a SBP ≥140 mmHg and/or DBP ≥90 mmHg and/or prescription of antihyper-tensive medication, as well as a SBP ≥130 mmHg and/or DBP ≥85 mmHg. ‘Not lean’ populations were defined by a mean population BMI of ≥23 kg/m2 and ‘lean’ pop-ulations by a mean population BMI of <23 kg/m2. Statistical analysis Hypertension was the outcome variable used in this study. Studies were pooled according to whether salt (NaCl) and/or sodium intake had been assessed as a categorical variable or continuous variable. The pooled effect sizes (ESs) for studies were calculated based on the weighting of the number of participants in each study. We pooled studies according to whether salt intake was assessed as a continuous or a categorical variable. We report the ESs and 95% CIs for the impact of salt on the prevalence of hypertension. A ratio of the two odds ratios was used to detect statistical differences between urban and rural re-gions. A p value ≤0.05 was deemed statistically signifi-cant.

When we pooled categorical analyses for the meta-analyses, categories of salt were compared to the given reference category in each study. We further stratified the meta-analysis by sex when such data were available. A random effects model was used rather than a fixed model because of the expected heterogeneity across studies. In the random effects model we attempted to integrate the heterogeneity among studies into a pooled estimate using the method of DerSimonian and Laird.26 Publication bias was not assessed by conducting funnel plot asymmetry tests because these tests would lack sufficient power to identify bias in meta-analyses with fewer than 10 studies identified.27 Sensitivity analyses were conducted by re-moving each study from the meta-analysis and comparing the point estimates before and after each study was ex-cluded. Analyses were performed using STATA Version 11.2 (Stata Corporation, College Station, Texas). RESULTS Search results and study characteristics Guidelines from MOOSE28 and PRISMA29 were followed for this review and meta-analyses. Of the 1567 publica-tions that were identified, 1066 remained after duplicates were removed (Figure 1). After excluding articles by re-viewing titles and abstracts, 161 full text articles were assessed for eligibility. After assessing the data presented in articles that met the eligibility criteria, one study was

404 AK Subasinghe, S Arabshahi, D Busingye, RG Evans, KZ Walker, MA Riddell and AG Thrift

further excluded because the authors failed to provide information about whether or not they had log trans-formed the variable of salt, nor did they explain which variables were adjusted for in the multivariable logistic model.30 Subsequently, 18 studies,31-48 in which the asso-ciation of salt intake with hypertension had been assessed in LMICs, were included in this review. Eleven of these studies were conducted in rural populations, six were conducted in urban populations, and one was conducted in both rural and urban populations. Overall, these 18 studies, with sample sizes ranging from 255 to 36,154, provided data for a total of 134,916 participants (Table 1).

Individual consumption of salt was estimated by vari-ous methods between the studies (Table 2), including reports of the total amount of salt consumed by a family for one year;32-35,47 by a direct question about salt added to food;36,37,45,46,48 or by a dietary method (seven day food records,38,40,43,44 food frequency questionnaire (FFQ),42 household questionnaire,21 or 24 hour recall).39,41 In addi-tion, different characteristics were adjusted for in the mul-tivariable logistic regression models (Table 2). Hypertension The mean prevalence of hypertension was reported for the total population in the majority of studies.33-37,39-46,48 From these estimates, the prevalence of hypertension ranged from 7.2% to 42% (Table 1). The greatest preva-lence of hypertension reported in urban populations was 36.7% and in rural populations it was 42%. Salt intake Estimates of mean daily salt intake were reported for nine populations, and ranged from 6.9 g/day to 42.3 g/day. The average salt and sodium intake reported in all of these populations were greater than the daily allowance rec-ommended (RDA) by the WHO (<5 g/day).49 The prev-alence of high salt intake ranged from 21.3% to 89.5% in

both rural and urban populations (Figure 2). The greatest prevalence of high salt intake reported was in urban India (87.8%)43 and in rural Nepal (89.5%).45

Salt intake and hypertension Studies in which salt was assessed as a categorical var-iable We were unable to detect a significant difference in the relationship between high salt intake and hypertension in studies conducted in rural populations (n=8) (pooled ES 1.36 95% CI 1.24, 1.48) and urban populations (n=3) (pooled ES 1.28 95% CI 1.13, 1.45, p for difference= 0.34, Figure 3). Studies in which salt was assessed as a continuous var-iable When salt was analysed continuously there was a greater association between salt intake and hypertension in urban (pooled ES 1.42 95% CI 1.19, 1.69) per one gram in-crease in salt (Figure 4) than in rural populations (pooled ES 1.07 95% CI 1.04, 1.10, p for difference <0.001). BMI, salt intake and hypertension The mean BMI for rural populations ranged from 19.5 kg/m2 to 24 kg/m2 and between 22.4 kg/m2 and 32.5 kg/m2 for urban populations. We found a greater impact of salt on hypertension in lean rural populations than non-lean rural populations when salt was analysed as a con-tinuous variable (p for difference <0.001, Figure 5). We were unable to assess the impact of salt on hypertension according to BMI in studies conducted in urban popula-tions as few investigators reported measures of BMI. Sensitivity analyses In sensitivity analyses, when we removed each study in-dividually, no single study significantly changed the pooled effect sizes for high salt intake and hypertension

Figure 1. Flowchart of search strategy and selection process


Table 1. Summary of studies where multivariable analysis was used to assess the relationship between salt and hypertension in rural and urban populations of low to middle income countries

Location Survey year

Sample size (n)

Age range (year)

Mean age (SD)

Mean BMI

Prevalence of hypertension (%) Mean blood pressure mmHg (SD) Total Men Women SBP DBP

Rural populations of LMICs Nandan County, China39 NR Bai Ku Yao 485 40-85 54 (11) 22.2 21.9 25.3 17.7 123 (18) 77 (10) Han Chinese 501 40-89 55 (11) 22.7 28.9 33.7 23.8 127 (18) 79 (11) Fuxin, Liaoning, China33 2004-06 ≥35 51.2 (11.8) Mongolian 9236 23.1 42 41.7 42.2 136 (23.2) 83.7 (12.8) Han 36,154 23.6 36.7 35.8 37.7 134 (22.6) 82.5 (12.7) Hubei, China 37 NR 3054 >35 51.4 (12.2) 22.1 26.9 – – 127 (20.2) 77.2 (11.4) Chirai Goan, Varanasi, India34 NR 316 ≥30 – 4.7† 7.2 5.6 8.8 Assam, India36 NR 3180 ≥30 – 10.2† 33.3 33.2 33.4 Rishi Valley, South India31 2006 1479 ≥18 39.7 (15.6) 19.5 11.4 16.2 7.3 115 (114, 116) 70.9 (70.3, 71.5) Moradabad, India38 NR 255 60-84 21.5 13 16 10 Fuxin County, Liaoning, China35 2004-06 Mongolian 8320 ≥35 50.4 (11.6) 23.5 11.6 – – Han 33,036 ≥35 50.6 (11.7) 23.0 14.7 – – Liaoning Province, China32 2004-06 23,178F ≥35 – 24 – – 38.6 Bhadrabas, Nepal45 2006 1218 ≥21 – 33.8 38.3 30.8 Moradabad, India 44 NR 1769 25-64 11† 17.7 17.4 17.9 Nanning, China 48 NR 2036 – ≥18 – 11.6 12.5 11 – – Urban populations of LMICs Trivandrum, India40 NR 1497 25-64 – 24.2 34.6 30.7 Hyderabad, India41 NR 165 30-50 – Moradabad, Trivandrum, Calcutta, Nagpur, Bombay, India43

3212F 25-64 23.2 25.6 125 (17) 82 (11)

Moradabad, India38 NR 311 60-84 22.4 – 58 48 Chennai, India 42‡ NR 1902 ≥20 25 West Bengal, India46 NR 1609 >18 – 32.5 Lomé, Togo47 2011 2002 36.7 34.6 38.4 LMICs: low-and-middle income countries; F: female; NR: not reported; SBP: systolic blood pressure; DBP: diastolic blood pressure. †Overweight BMI ≥23 kg/m2; BMI >25 kg/m2. ‡>130/85 mmHg SBP/DBP.


Table 2. Studies of the association between salt intake and hypertension in rural and urban areas of LMICs: multivariable analyses

Location Salt intake OR (95% CI) Total Adjustments Prevalence of

high salt intake Individual daily salt intake calculated from

SBP ≥140 mmHg and/or DBP ≥90 mmHg or using antihypertensive medication China33

Mongolian Han

Cont. g/d per increase in one gram

1.01 (1.00, 1.01) 1.01 (1.00, 1.01)

Age, sex, BMI, WC, education, smoking status, alcohol consumption, lipid disorder, diabetes, and FHX HT

Total amount of salt consumed by a family for 1 year was used to calculate individual consumption

China32 <14 g/d ≥14 g/d

1.0 1.2 (1.1-1.3)

Age, ethnicity, education, income, smoking status, drinking status, BMI, WC, lipid disorder, diabetes, and FHX HT

Total amount of salt consumed by a family for 1 year was used to calculate individual consumption

China37 ≤6 g/d >6 g/d

1.0 1.15 (1.05, 1.25)

BMI, educational level, predominantly sedentary work, physical exercise, alcohol consumption, positive FHX HT, vegetable and fruit intake, animal insides intake, and dysart-eriotony (pressure difference >60 mmHg or 20 mmHg be-tween SBP and DBP), where appropriate

Dietary questionnaire. High salt intake defined as >6 g/d

India31 ≤ median intake > median intake

1.0 2.39 (1.36, 4.21) M 0.85 (0.44, 1.64) F

Model 2: M: age, BMI, alcohol consumption, physical activity, education, and cholesterol F: age, BMI, smokeless tobacco, education, and triglyceride

Salt added to household cooking divided by the number of individuals in each family

SBP ≥140 mmHg and/or DBP ≥90 mmHg Moradabad, India38 <8 g/d

≥8 g/d

1.18 (1.12, 1.27) M 1.17 (1.12, 1.26) F

Age, BMI, WHR, SES, energy expenditure, 2 hour plasma insulin, and 2 hour blood glucose

54.1% Seven day food intake records were obtained for all individuals in the household.

Assam, India36† No extra salt intake Extra salt intake ≥ ¼ teaspoon

1.0 1.45 (1.21, 1.73) 1.36 (1.05, 1.77) M 1.51 (1.19, 1.92) F

Age, sex, BMI, WHR, type of work, smoking status, tobacco chewing, alcohol consumption, and marital status

50.1% Direct question- daily intake ≥¼ teaspoon of salt or ≥½ teaspoon salt/d

China35‡ Mongolian <13.5 g/d

≥13.5 g/d 1.0 1.42 (1.26, 1.60)

Age, sex, BMI, smoking status, alcohol consumption, and FHX HT

Total amount of salt consumed by a family for one year to calculate individual consumption

Han <14.6 g/d ≥14.6 g/d

1.0 1.31 (1.22, 1.40)

Age, sex, BMI, smoking status, alcohol consumption, and FHX HT

Total amount of salt consumed by a family for one year to calculate individual consumption

China39§ Bai Ku Yao Han


1.55 (1.30, 2.94) 1.57 (1.22, 2.56)

Age, BMI, WC, education level, physical activity, triglycerides, total energy, total fat, total dietary fibre

24 hour recall questionnaire

LMIC’s: low to middle income countries; Cont: continuous; d: day; Ref: reference; M: males; F: females; BP: blood pressure; SBP: systolic blood pressure; DBP: diastolic blood pressure; BMI: body mass index; FFQ: food frequency questionnaire; WHR: waist/hip ratio; WC: waist circumference; SES: socioeconomic status; FHX HT: family history of hypertension. †Participants who used antihypertensive medication were excluded from this analysis.

‡Participants did not take antihypertensive medications the last two weeks before they were examined. §Sodium intake. ¶Odds ratio for the third tertile.


Table 2. Studies of the association between salt intake and hypertension in rural and urban areas of LMICs: multivariable analyses (cont.) Location Salt intake OR (95% CI)

Total Adjustments Prevalence of

high salt intake Individual daily salt intake calculated from

Bhadrabas, Nepal45 <5 g/d ≥5 g/d

1.0 1.54 (1.0-2.35)

Age, sex, physical activity, tobacco, high waist circumference, BMI

89.5% Visual aids (sachets of salt) were showed to participants to estimate amount of daily salt consumption

Varanasi, India34 ≤75 g/d >75 g/d

1.0 2.68 (0.09-6.32)

BMI, fat intake, and energy intake 32.6%

Direct questions: how much salt was added during cooking, how much salt was bought and consumed by the household each month

Nanning, China48 Not a high salt diet High salt diet

1.0 1.48 (1.04-2.10)

21.3% Direct questions: Do you consume a high salt diet?

Moradabad, India44 Cont. g/d per increase in one gram

1.0 1.31 (1.06-1.68) M 1.28 (1.20-1.46) F

Age 68% Seven day food records

Urban Moradabad, India38 ≤8 g/d

>8 g/d 1.0 1.45 (1.15-1.65) M 1.47 (1.35-1.58) F

Age, BMI, energy expenditure, plasma insulin, blood glucose, SES

48.8% Seven day food records

Trivandrum, India40 ≤8 g/d >8 g/d

1.0 1.35 (1.02-1.65) M 1.38 (1.04-1.57) F

Age 38.2% Seven day food records

Hyderabad, India41 Cont. g/d per increase in one gram

4.36¶ (1.11-17.12) M 3.43¶ (1.23-9.48) F

Energy, protein, fat, visible fat, energy fat and sodium: potassium ratio

24 hour recall method

Moradabad, Trivandrum, Calcutta, Nagpur, Bombay, India43


1.45 (1.12-1.85) Age, BMI, overweight, sedentary lifestyle 87.8% Seven day food records

West Bengal, India46 No extra salt added to meal Extra salt added to meal

1.0 1.15 (0.87-1.50)

Age, occupation, BMI, smoking, diet, ischaemic heart disease

26.4% Direct questions: how much salt was added during cooking and extra salt added to the meal

Lomé, Togo47 Moderate consumption of salt High consumption of salt

1.0 1.12 (1.07-1.98)

Age, sex 10 minute questionnaire

SBP ≥130 mmHg and/or DBP ≥85 mmHg Chennai, India42 Cont. g/d

per increase in one gram 1.16 (1.12-1.21) Age, gender, BMI, total energy and dietary

fat intake FFQ

LMIC’s: low to middle income countries; Cont: continuous; d: day; Ref: reference; M: males; F: females; BP: blood pressure; SBP: systolic blood pressure; DBP: diastolic blood pressure; BMI: body mass index; FFQ: food frequency questionnaire; WHR: waist/hip ratio; WC: waist circumference; SES: socioeconomic status; FHX HT: family history of hypertension. †Participants who used antihypertensive medication were excluded from this analysis.

‡Participants did not take antihypertensive medications the last two weeks before they were examined. §Sodium intake. ¶Odds ratio for the third tertile.


(data not shown). DISCUSSION From our meta-analysis of studies including 134,916 par-ticipants, we found that high salt intake was prevalent in both rural and urban populations of LMICs. The preva-lence of high salt intake (26.4-89.5%) and hypertension (7.2-42%) reported in urban and rural populations of LMICs ranged widely. Our findings indicate that the

prevalence of high salt intake in our study may even be as high as that has been reported in the UK and the US (77-95.4%).50-51 Furthermore, we have shown that excess salt intake is associated with hypertension irrespective of whether populations are urban or rural.

We found that the prevalence of excessive salt intake was high in both rural and urban populations of LMICs. It has been argued that evidence for high salt intake in ur-ban populations of LMICs further reflects their nutritional

Figure 4. Meta-analysis of the association between salt intake as a continuous variable and hypertension in studies conducted in rural (n=4) 33,38,39,44 and urban populations (n=4).38,41-43 ES: effect size; CI: confidence interval; I-squared: test for heterogeneity; M: male; F: female.

Figure 5. Meta-analysis of studies in which the association between salt intake and hypertension was assessed in ‘not lean’ rural populations33 compared with ‘lean’ rural populations (n=2).38-39 ES: effect size; CI: confidence interval; I-squared: test for heteroge-neity.

Figure 2. Prevalence of high salt intake in studies of hypertension in rural and urban populations (n=9)34,36,38,40,43-46,48

Figure 3. Meta-analysis of the association between salt intake (categorised high/low) and hypertension in studies conducted in rural populations (n= 8)31,34-37,45,48 and urban populations (n=3).40,46,47 ES: effect size; CI: confidence interval; I-squared: test for hetero-geneity; M: male; F: female.


transition to ‘Western-style’ diets.52 This nutritional tran-sition is characterised by the increase in availability of food and adoption of westernised dietary habits such as a greater consumption of processed foods, fats, sodium, and sugars.22 It might, therefore, seem surprising that salt in-take is so high in rural populations of LMICs given the relative lack of processed foods documented in these are-as in previous population based studies.53-55 Possible ex-planations for this phenomenon include the fact that peo-ple who labour physically in hot climates require more salt than those who are not exposed to such conditions.56 Salt is often still used as an important food preservative in communities with little or no access to refrigeration or other methods for safe storage and transport of food.57

Additionally, limited available data indicate that most salt in rural India is added during cooking, and/or at the table, compared to urban populations and high income countries where processed foods largely contribute to overall popu-lation salt intake.58 The high salt intake evident in rural populations of LMICs indicates that here, as in high in-come countries, the implementation of public health pro-grams is germane.11 However, importantly, the messages delivered by such programs should be tailored to the un-derlying causes of excessive salt intake in each popula-tion.

Because the intake of salt in urban and rural areas may be derived from different sources, interventions to reduce salt intake at a population level should be developed to

Figure 4. Meta-analysis of the association between salt intake as a continuous variable and hypertension from rural (n=4)33,38,39,44 and urban populations (n=4).38,41-43 ES: effect size; CI: confidence interval; I-squared: test for heterogeneity; M: male; F: female.

Figure 5. Populations in which the association between salt intake and hypertension was assessed in ‘not lean’ rural populations33 com-pared with ‘lean’ rural populations (n=3).38-39 ES: effect size; CI: confidence interval; I-squared: test for heterogeneity.


target these populations separately. In urban areas, gov-ernment initiatives aimed at developing greater individual awareness of salt intake, or public nutritional education campaigns may be successful in triggering a reduction in consumption of processed foods. Government run media campaigns, along with the provision of clear labelling of sodium content on the packaging of processed foods, are some of the approaches that have been successfully im-plemented in over ten countries in Europe and Asia.59 In rural areas, initiatives to reduce salt intake should be im-plemented at a community level in order to galvanise a change in dietary behaviour. For example, local health-care practitioners could work together with communities to encourage restricting the amount of salt added to meals. If nutritional interventions are developed at a population level in urban areas, and at a grassroots level in rural are-as, there would be a greater likelihood of reduced salt intake in LMICs.

We found a greater impact of salt on hypertension in urban than rural populations when salt was analysed con-tinuously. However, no differences were detected be-tween the two populations when salt was analysed in a dichotomous fashion. A potential explanation for these disparate findings lies in the heterogeneity of definitions used to categorise high salt intake in categorical analyses. These definitions of high salt intake varied in the cut-point used to separate categories and in the units used. For example >6 grams per day was used to define high salt intake in one study, and >75 grams per day in another study. This means that for categorical analyses, the asso-ciation between salt and hypertension was quantified in a different way in each study and so combined estimates would be expected to be heterogeneous. This heterogenei-ty may have further reduced the likelihood of finding a difference between urban and rural regions. In contrast, when salt was analysed continuously, our meta-analysis provides a more robust interpretation of pooled effect sizes because the same quantity of one gram of salt was used.

A noteworthy observation is that when salt was ana-lysed continuously, in studies conducted in rural regions, we found that the association between salt and hyperten-sion was greater in lean than in non-lean populations. This finding contradicts our hypothesis that the effect of BMI on the relationship between salt and hypertension would be more pronounced in non-lean than in lean popu-lations. Given that lean populations have been accepted as a proxy for maternal malnutrition,60 the prevalence of maternal malnutrition may be greater in the lean popula-tions than in not-lean populations included in this meta-analysis. Maternal malnutrition appears to impede devel-opment of the fetal kidney, leading to deficient nephron endowment.21,61 This deficient nephron endowment can pre-dispose individuals to the development of hyperten-sion in adulthood, at least in part, by rendering their arte-rial pressure salt-sensitive.21 This mechanism may pro-vide an explanation for our findings for a greater effect of salt on hypertension in lean than in non-lean populations. However, this proposition is only speculative, as this hy-pothesis could only be tested in a prospective fashion.

A major limitation to our study was that data on salt in-take were obtained from different methods of assessments

in each of the studies included. Moreover, the majority of studies included the assessment of salt using self-report only. Because the validity of salt intake assessed using this method is highly questionable,62 the estimates of salt intake reported in these studies may not be accurate. In addition, different definitions of hypertension were used in each study, and this may have introduced further heter-ogeneity in our analyses when salt was analysed categori-cally. Finally, there were only 18 population-based stud-ies that met the eligibility criteria, revealing a considera-ble gap in the literature related to associations between salt and hypertension in LMICs.

Strengths to this review were that we only included studies in which multivariable models were used to assess independent associations between salt and hypertension and that the outcome of hypertension was directly meas-ured in each study. Furthermore, to our knowledge no other systematic review and or meta-analyses have been conducted to investigate the relationship between salt and hypertension in rural and urban populations of LMICs, and so this summary addresses an important knowledge-gap. Conclusion In our meta-analysis of pooled results from 18 population based studies we found that high salt intake is very preva-lent in rural and urban populations of LMICs. Further-more, the impact of salt intake on hypertension appeared to be greater in urban than in rural regions of LMICs. Additionally, we found that salt had a greater impact on hypertension in lean compared with non-lean populations in rural regions. Public health initiatives are required to reduce salt intake in both rural and urban populations to potentially reduce the burden of hypertension in these regions. FUNDING AGT was supported by a Senior Research Fellowship from the National Health & Medical Research Council (NHMRC), Aus-tralia (1042600). MAR and SA were supported through project grants from the NHMRC (1005740 and 1040030). DB was sup-ported by a Monash Faculty of Medicine International Post-graduate Scholarship and a Monash Graduate Scholarship. AKS was supported by a Monash Graduate Scholarship. AUTHOR DISCLOSURES All authors declared no conflict of interest. REFERENCES 1. Ibrahim MM, Damasceno A. Hypertension in developing

countries. Lancet. 2012;380:611-9. doi: 10.1016/s0140-673 6(12)60861-7.

2. World Health Organization. WHO Global Report: preventing chronic disease: a vital investment. Geneva: World Health Organization; 2005.

3. Bansal SK, Saxena V, Kandpal SD, Gray WK, Walker RW, Goel D. The prevalence of hypertension and hypertension risk factors in a rural Indian community: a prospective door-to-door study. J Cardiovasc Dis Res. 2012;3:117-23. doi: 10. 4103/0975-3583.95365.

4. Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: general considerations, the epidemiologic transition, risk factors, and impact of


urbanization. Circulation. 2001;104:2746-53. doi: 10.1161/ hc4601.099487.

5. Pereira M, Lunet N, Azevedo A, Barros H. Differences in prevalence, awareness, treatment and control of hypertension between developing and developed countries. J Hypertens. 2009;27:963-75. doi: 10.1097/HJH.0b013e3283 282f65.

6. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global burden of hypertension: analysis of worldwide data. Lancet. 2005;365:217-23. doi: 10.1016/s01 40-6736(05)17741-1.

7. Ajay VS, Gupta R, Panniyammakkal J, Chaturvedi V, Prabhakaran D, Reddy KS. National cardiovascular disease database. India: Ministry of Health & Family, Government of India and WHO; 2005.

8. Drenjancevic-Peric I, Jelakovic B, Lombard JH, Kunert MP, Kibel A, Gros M. High-salt diet and hypertension: focus on the Renin-Angiotensin system. Kidney Blood Press Res. 2011;34:1-11. doi: 10.1159/000320387.

9. Radhika G, Sathya RM, Sudha V, Ganesan A, Mohan V. Dietary salt intake and hypertension in an urban south Indian population--[CURES - 53]. J Assoc Physicians India. 2007; 55:405-11. doi: 10.1017/S136898001000203X.

10. Frisoli TM, Schmieder RE, Grodzicki T, Messerli FH. Salt and hypertension: is salt dietary reduction worth the effort? Am J Med. 2012;125:433-9. doi: 10.1016/j.amjmed.2011.10. 023.

11. Gupta R, Guptha S. Strategies for initial management of hypertension. Indian Journal of Medical Research. 2010; 132:531-42.

12. Appel LJ, Brands MW, Daniels SR, Karanja N, Elmer PJ, Sacks FM. Dietary approaches to prevent and treat hypertension: a scientific statement from the American Heart Association. Hypertension. 2006;47:296-308. doi: 10. 1161/ 01.hyp.0000202568.01167.b6.

13. Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, Bray GA, Harsha D et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344:3-10. doi: 10.1056/nejm2001010 43440101.

14. Elliott P, Stamler J, Nichols R, Dyer AR, Stamler R, Kesteloot H et al. Intersalt revisited: further analyses of 24 hour sodium excretion and blood pressure within and across populations. Intersalt Cooperative Research Group. BMJ. 1996;312:1249-53. doi: 10.1136/BMJ.312.7041.1249.

15. Intersalt: an international study of electrolyte excretion and blood pressure. Results for 24 hour urinary sodium and potassium excretion. Intersalt Cooperative Research Group. BMJ. 1988;297:319-28. doi: 10.1136/BMJ.297.6644.319.

16. He FJ, Li J, Macgregor GA. Effect of longer term modest salt reduction on blood pressure: Cochrane systematic review and meta-analysis of randomised trials. BMJ. 2013; 346:f1325. doi: 10.1136/bmj.f1325.

17. Dyer AR, Elliott P, Shipley M, Stamler R, Stamler J. Body mass index and associations of sodium and potassium with blood pressure in INTERSALT. Hypertension. 1994;23:729-36. doi: 10.1161/01.HYP.23.6.729.

18. Rocchini AP, Key J, Bondie D, Chico R, Moorehead C, Katch V et al. The effect of weight loss on the sensitivity of blood pressure to sodium in obese adolescents. New Engl J Med. 1989;321:580-5. doi: 10.1056/nejm198908313210905.

19. Ramachandran A, Snehalatha C. Rising burden of obesity in Asia. J Obes. 2010;2010. pii: 868573. doi: 10.1155/2010/86 8573.

20. Stamler J. The INTERSALT Study: background, methods, findings, and implications. Am J Clin Nutr. 1997;65(2

Suppl):626S-42S. doi: 10.1161/01.HYP.14.5.570. 21. Thrift AG, Srikanth V, Fitzgerald SM, Kalyanram K, Kartik

K, Hoppe CC et al. Potential roles of high salt intake and maternal malnutrition in the development of hypertension in disadvantaged populations. Clin Exp Pharmacol Physiol. 2010;37:e78-90. doi: 10.1111/j.1440-1681.2009.05266.x.

22. Popkin BM. The nutrition transition and obesity in the developing world. J Nutr. 2001;131:871S-3S. doi: 10.1111/ j.1467-8659.2003.00225.x.

23. World Health Organization. Reducing salt intake in populations. Paris, France: World Health Organization; 2006.

24. Borzecki AM, Oliveria SA, Berlowitz DR. Barriers to hypertension control. Am Heart J. 2005;149:785-94. doi: 10. 1016/j.ahj.2005.01.047.

25. World Bank. World development indicators. World Bank, Washington, DC; 2012. [cited 2014/04/07]; Available from: http://data.worldbank.org/about/country-classifications/coun try-and-lending-groups.

26. Hedges LV, Olkin I. Statistical methods for meta-analysis. San Diego: Academic Press; 1985.

27. Higgins J, Green, S. Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]. The Cochrane Collaboration: 2011. Available from: www. cochrane-handbook.org.

28. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008-12. doi: 10.1001/jama.283.15.2008

29. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336-41. doi: 10. 1016/j.ijsu.2010.02.007.

30. Guerrero-Romero F, Rodriguez-Moran M, Sandoval-Herrrera F, Alvarado-Ruiz R. Prevalence of hypertension in indigenous inhabitants of traditional communities from the north of Mexico. J Hum Hypertens. 2000;14:555-9. doi: 10. 1038/sj.jhh.1001067.

31. Thrift AG, Evans RG, Kalyanram K, Kartik K, Fitzgerald SM, Srikanth V. Gender-specific effects of caste and salt on hypertension in poverty: a population-based study. J Hypertens. 2011;29:443-50. doi: 10.1097/HJH.0b013e32834 1888c.

32. Sun Z, Zheng L, Detrano R, Zhang D, Zhang X, Xu C et al. The accelerating epidemic of hypertension among rural Chinese women: results from Liaoning Province. Am J Hypertens. 2008;21:784-8. doi: 10.1038/ajh.2008.170.

33. Sun Z, Zheng L, Xu C, Li J, Zhang X, Liu S et al. Prevalence of prehypertension, hypertension and, associated risk factors in Mongolian and Han Chinese populations in Northeast China. Int J Cardiol. 2008;128:250-4. doi: 10. 1016/j.ijcard.2007.08.127.

34. Goel NK, Kaur P. Role of various risk factors in the epidemiology of hypertension in a rural community of Varanasi district. Indian J Public Health. 1996;40:71-6.

35. Li J, Xu C, Sun Z, Zheng L, Zhang D, Zhang X et al. Prevalence and risk factors for isolated untreated systolic hypertension in rural Mongolian and Han populations. Acta Cardiol. 2008;63:389-93. doi: 10.2143/AC.63.3.1020317.

36. Hazarika NC, Narain K, Biswas D, Kalita HC, Mahanta J. Hypertension in the native rural population of Assam. Natl Med J India. 2004;17:300-4.

37. Huang S, Xu Y, Yue L, Wei S, Liu L, Gan X et al. Evaluating the risk of hypertension using an artificial neural network method in rural residents over the age of 35 years in a Chinese area. Hypertens Res. 2010;33:722-6. doi: 10.1038/ hr.2010.73.


38. Singh RB, Rastogi SS, Rastogi V, Niaz MA, Madhu SV, Chen M et al. Blood pressure trends, plasma insulin levels and risk factors in rural and urban elderly populations of north India. Coron Artery Dis. 1997;8:463-8. doi: 10.1097/ 00019501-199707000-00009.

39. Ruixing Y, Weixiong L, Hanjun Y, Dezhai Y, Shuquan L, Shangling P et al. Diet, lifestyle, and blood pressure of the middle-aged and elderly in the Guangxi Bai Ku Yao and Han populations. Am J Hypertens. 2008;21:382-7. doi: 10. 1038/ ajh.2008.1.

40. Beegom R, Singh RB. Association of higher saturated fat intake with higher risk of hypertension in an urban population of Trivandrum in South India. Int J Cardiol. 1997; 58:63-70. doi: 10.1016/s0167-5273(96)02842-2.

41. Kodali V. Dietary factors as determinants of hypertension: A case control study in an urban Indian population. Asia Pac J Clin Nutr. 1999;8:184-9. doi: 10.1046/j.1440-6047.1999.00 092.

42. Radhika G, Sathya RM, Sudha V, Ganesan A, Mohan V. Dietary salt intake and hypertension in an urban South Indian population - [CURES - 53]. J Assoc Physicians India. 2007;55:405-11.

43. Singh RB, Beegom R, Mehta AS, Niaz MA, De AK, Haque M et al. Prevalence and risk factors of hypertension and age-specific blood pressures in five cities: a study of Indian women. Int J Cardiol. 1998;63:165-73. doi: 10.1016/s0167-5273(97)00296-9.

44. Singh RB, Rastogi V, Niaz MA, Sharma JP, Raghuvanshi R, Moshira M. Epidemiological study of magnesium status and risk of hypertension in a rural population of north India. Magnes Res. 1996;9:173-81.

45. Vaidya A, Pathak RP, Pandey MR. Prevalence of hypertension in Nepalese community triples in 25 years: A repeat cross-sectional study in rural Kathmandu. Indian Heart J. 2012;64:128-31. doi: 10.1016/S0019-4832(12)600 45-5.

46. Das SK, Sanyal K, Basu A. Study of urban community survey in India: growing trend of high prevalence of hypertension in a developing country. Int J Med Sci. 2005; 2:70-8. doi: 10.7150/ijms.2.70.

47. Yayehd K, Damorou F, Akakpo R, Tchérou T, N'Da NW, Pessinaba S et al. Prevalence and determinants of hypertension in Lomé (Togo): Results of a screening in May 2011. Ann Cardiol Angeiol. 2013;62:43-50. doi: 10.1016/j. ancard.2012.09.006.

48. Pan L, Song YS, Ma R, Ye XD, Liao YH. An epidemiology study on hypertension in the Zhuang nationality village from the Guangxi Zhuang Autonomous Region. Zhonghua xin xue guan bing za zhi. 2013;41:790-4. (In Chinese)

49. World Health Organization. Diet, Nutrition and Prevention of Chronic Diseases. Geneva: World Health Organization;

2003. doi: 10.1111/j.1753-4887.1991.tb07370.x. 50. Department of Health and the Food Standards Agency.

National Diet and Nutrition Survey- Headline results from Years 1 and 2 (combined) of the rolling programme. 2008-2009. United Kingdom: Department of Health and the Food Standards Agency; 2010.

51. Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey: trends in the prevalence of excess dietary sodium intake-United States. 2003-2010. Atlanta: U.S Department of Health and Human Services; 2013.

52. Popkin BM. Global changes in diet and activity patterns as drivers of the nutrition transition. Nestle Nutr Workshop Ser Pediatr Program. 2009;63:1-10; discussion-4, 259-68. doi: 10.1159/000209967.

53. Vedantam A, Subramanian V, Rao NV, John KR. Malnutrition in free-living elderly in rural south India: prevalence and risk factors. Public Health Nutr. 2010;13: 1328-32. doi: 10.1017/s1368980009991674.

54. Vijayaraghavan K, Rao DH. Diet & nutrition situation in rural India. Indian J Med Res. 1998;108:243-53.

55. Zhang M, Binns CW, Lee AH. Dietary patterns and nutrient intake of adult women in south-east China: a nutrition study in Zhejiang province. Asia Pac J Clin Nutr. 2002;11:13-21. doi: 10.1046/j.1440-6047.2002.00259.x.

56. Rao SP. Salt intake among Indians: role in non-communicable diseases. Medical Journal of Western India. 2013;41;7.

57. Indian Council of Medical Research. Nutrient Requirements and Recommended Dietary Allowances For Indians. Hyderabad: National Institute of Nutrition; 2009.

58. Mohan S, Prabhakaran, D. Review of salt and health: situation in south-east Asia. South-East Asia Regional Office: World Health Organization; 2013

59. European Commission Directorate. Implementation of the EU salt reduction framework: results of member states survey. Luxembourg: Publications Office of the European Union: 2012

60. Arabshahi S, Busingye D, Subasinghe A, Evans R, Riddell M, Thrift A. Adiposity has a greater impact on hypertension in lean than not-lean populations: a systematic review and meta-analysis. Eur J Epidemiol. 2014:29:311-24. doi: 10. 1007/s10654-014-9911-6.

61. Barker DJ BS. Developmental antecedents of cardiovascular disease: a historical perspective. J Am Soc Nephrol. 2005; 16:2537-44. doi: 10.1681/ASN.2005020160.

62. Bingham SA, Gill C, Welch A, Day K, Cassidy A, Khaw KT et al. Comparison of dietary assessment methods in nutritional epidemiology: weighed records v. 24 h recalls, food-frequency questionnaires and estimated-diet records. Br J Nutr. 1994;72:619-43. doi: 10.1079/BJN19940064.


Original Article Association between salt and hypertension in rural and urban populations of low to middle income countries: a systematic review and meta-analysis of population based studies Asvini K Subasinghe PhD1, Simin Arabshahi PhD1, Doreen Busingye PhD1, Roger G Evans PhD2, Karen Z Walker PhD3, Michaela A Riddell PhD1, Amanda G Thrift PhD1,4 1Department of Medicine, School of Clinical Sciences at Monash Health, Monash University, Melbourne, Australia 2Department of Physiology, Monash University, Melbourne, Australia 3Department of Nutrition and Dietetics, Monash University, Melbourne, Australia 4Florey Neuroscience Institutes, Heidelberg, Victoria, Australia

低中收入国家城乡居民盐摄入量与高血压的关系：以人

群为基础的系统性研究和 meta 分析

背景与目的：在全球高血压患病率对死亡率贡献最大，在低中等收入国家高血

压的患病率呈增加的趋势。在低中收入国家的城市，过量的盐摄入与高血压患

病风险的增加有关，我们旨在研究农村地区是否亦如此。方法与研究设计：使

用多变量模型，我们进行了低中收入国家盐与高血压关系的 meta 分析。结果：

共分析了来自 18 个研究的 134,916 名参与者。农村和城市居民高盐摄入的比例

在 21.3%到 89.5%之间。当盐摄入量作为连续型变量分析时，发现城市地区

（n=4） [合并效应尺度（ES）为 1.42，95% CI 为 1.19-1.69]比农村地区

（n=4）[合并效应尺度（ES）为 1.07，95% CI 为 1.04-1.10，p<0.001]。当盐摄

入量作为连续型变量分析时，体型偏瘦的农村人群（BMI <23 kg/m2）比体型不

偏瘦的农村人群（BMI ≥23 kg/m2，p<0.001）盐摄入对高血压患病的影响更

大。结论：农村和城市地区高盐摄入的比例相近。与农村居民相比，城市居民

过多的盐摄入对高血压患病率的影响更大。在农村居民中，BMI 能调节盐摄入

和高血压的关系。

关键词：高血压、盐、城市、农村、meta 分析

association between salt and hypertension in rural and urban...

Documents